CN112098848B - Battery discharge capacity prediction method and system and readable storage medium - Google Patents
Battery discharge capacity prediction method and system and readable storage medium Download PDFInfo
- Publication number
- CN112098848B CN112098848B CN202010960776.3A CN202010960776A CN112098848B CN 112098848 B CN112098848 B CN 112098848B CN 202010960776 A CN202010960776 A CN 202010960776A CN 112098848 B CN112098848 B CN 112098848B
- Authority
- CN
- China
- Prior art keywords
- voltage
- battery
- discharge capacity
- capacity
- sequence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
- G01R31/385—Arrangements for measuring battery or accumulator variables
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
- Y02E60/10—Energy storage using batteries
Abstract
The invention relates to the technical field of batteries, in particular to a method and a system for predicting the discharge capacity of a battery and a readable storage medium, wherein the method comprises the steps of obtaining the charging voltage and the discharge capacity of a plurality of time points with equal intervals in the charging stage of the battery; processing the obtained charging voltage and the discharging capacity to obtain a plurality of discharging capacities corresponding to the voltage points at equal intervals, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the plurality of voltage points at equal intervals and the discharging capacities corresponding to the voltage points at equal intervals; performing characteristic extraction on the first incremental capacity curve to obtain a charging voltage when the battery is charged and a characteristic parameter related to the corresponding discharge capacity; and inputting the obtained characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery. According to the embodiment, the discharge capacity of the battery is predicted according to the extracted characteristic parameters during the charge and discharge of the battery by adopting the pre-trained prediction model, so that the accuracy of prediction of the discharge capacity of the battery is improved.
Description
Technical Field
The invention relates to the technical field of batteries, in particular to a method and a system for predicting battery discharge capacity and a readable storage medium.
Background
With the vigorous promotion of the new energy automobile industry by the policy of China and the rapid development of the new energy automobile technology of China, the market volume of new energy automobiles of China is huge and the world is ahead, and China becomes the first big country of new energy automobiles.
The power battery is the most core part of the new energy automobile. Compared with the traditional battery, the lithium ion battery has the advantages of high energy density, long service life, no memory effect, good charge and discharge performance and the like, thereby being widely applied to new energy automobiles. Lithium ion batteries using lithium iron phosphate materials as the positive electrode materials have become the main flow power batteries used by various new energy automobile manufacturers due to the stable performance of the lithium ion batteries.
According to the relevant standards, when the capacity of the vehicle power battery is lower than 80% of the maximum capacity, the vehicle power battery faces the problems of performance reduction, safety reduction and the like, and is not suitable for being continuously used on new energy vehicles and faces retirement. With the development of new energy automobile industry in China, the retirement of power batteries becomes a large scale. If the power battery is directly discarded into the environment without being recycled after being retired, serious negative effects are caused to the environment. Therefore, relevant regulations have been issued in China to guide the waste power storage batteries to carry out multilevel and multipurpose reasonable utilization.
The cascade utilization of the retired power battery can be roughly divided into a whole package grade, a module grade cascade and a single battery part disassembly for direct utilization. Because the echelon utilization battery is used for a certain period, the problems of increased safety risk, increased capacity attenuation speed, increased battery faults, increased running cost and the like exist when the battery is continuously used, in order to realize safe and stable running of the echelon utilization battery, the health state of the echelon utilization battery needs to be evaluated, and the near-short-term discharge capacity of the running echelon utilization battery needs to be predicted to judge whether the situation that the battery capacity is about to rapidly attenuate occurs.
The existing battery state of health assessment and capacity prediction methods are mainly based on battery models. In practice, the battery has complicated and changeable surrounding environment in the working state, the cycle characteristics of the battery are influenced by various factors, so that many uncertain parameters exist in modeling, and the discharge capacity of the battery is not accurately predicted.
Disclosure of Invention
The invention mainly solves the technical problem that the prediction of the battery discharge capacity in the prior art is not accurate.
A method of predicting battery discharge capacity, comprising:
acquiring charging voltage and discharging capacity of a plurality of equal interval time points in a battery charging stage;
processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of equally spaced voltage points, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacities of the battery according to the equally spaced voltage points and the corresponding discharging capacities thereof;
extracting the characteristics of the first incremental capacity curve to obtain the charging voltage when the battery is charged and the corresponding characteristic parameters related to the discharge capacity;
and inputting the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery.
In one embodiment, after obtaining the discharge capacities corresponding to the plurality of equally spaced voltage points, the method further includes: and carrying out differential processing of the discharge capacity relative to the charge voltage on the plurality of voltage points and the corresponding discharge capacities thereof.
In one embodiment, after obtaining the first incremental capacity curve, the method further includes: and performing smooth noise reduction processing on the first incremental capacity curve by adopting a Savitzky-Golay filter.
In one embodiment, the obtaining the charging voltage and the discharging capacity at a plurality of equally spaced time points in the charging phase of the battery includes:
acquiring and recording an equal time interval sequence of a constant voltage charging stage in a charging period of K previous cycles of the battery from V1 to V2 and corresponding charging voltage and discharging capacity, wherein the equal time interval sequence comprises a plurality of equal interval time points, K is more than or equal to 1, and V2 is more than V1;
the step of processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of voltage points with equal intervals comprises the following steps:
dividing a voltage interval formed by the charging voltages V1-V2 into a first voltage sequence comprising a plurality of voltage points at equal intervals, and collecting the discharge capacity corresponding to each voltage point at equal intervals.
In one embodiment, the differentiating process of the discharge capacity with respect to the charge voltage for the plurality of voltage points and the corresponding discharge capacities thereof includes:
substituting the first voltage sequence into a preset linear spline interpolation function expression to obtain a first discharge capacity sequence;
and carrying out differential processing of the discharge capacity relative to the charge voltage on the first discharge capacity sequence and the first voltage sequence to obtain a second voltage sequence and a second discharge capacity sequence corresponding to the second voltage sequence.
In one embodiment, the obtaining a first incremental capacity curve representing a corresponding relationship between a charging voltage and a discharging capacity of a battery according to the plurality of equally spaced voltage points and the corresponding discharging capacities thereof includes: drawing a first incremental capacity curve by taking the second voltage sequence as a horizontal axis and the second discharge capacity sequence as a vertical axis;
the smoothing and denoising the first incremental capacity curve by using the Savitzky-Golay filter comprises:
and adopting a Savitzky-Golay filter to perform smooth noise reduction treatment by taking the second voltage sequence as an independent variable and the second discharge capacity sequence as a dependent variable to obtain a third discharge capacity sequence, and drawing a smooth first incremental capacity curve by taking the second voltage sequence as a horizontal axis and the third discharge capacity sequence as a vertical axis.
In an embodiment, after the characteristic parameter is input into a pre-trained prediction model to obtain the current discharge capacity of the battery, the method further includes training the prediction model by using the characteristic parameter as training data of the prediction model.
In one embodiment, during the training of the prediction model, a minimum mean square error of the prediction model is calculated, and the hyper-parameters of the prediction model are adjusted according to the minimum mean square error to optimize the prediction model.
A battery discharge capacity prediction system comprising:
the acquisition module is used for acquiring the charging voltage and the discharging capacity of a plurality of equal interval time points in the charging stage of the battery;
the first processing module is used for processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of equally spaced voltage points and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the equally spaced voltage points and the corresponding discharging capacities;
the extraction module is used for extracting the characteristics of the first incremental capacity curve to obtain the charging voltage when the battery is charged and the corresponding characteristic parameters related to the discharge capacity;
and the prediction module is used for inputting the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery.
A computer-readable storage medium comprising a program executable by a processor to implement the battery discharge capacity prediction method as described above.
The method for predicting the discharge capacity of the battery according to the embodiment comprises the following steps: acquiring charging voltage and discharging capacity of a plurality of equal interval time points in a battery charging stage; processing the obtained charging voltage and the obtained discharging capacitor to obtain discharging capacities corresponding to a plurality of equally spaced voltage points, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the equally spaced voltage points and the corresponding discharging capacities thereof; performing characteristic extraction on the first incremental capacity curve to obtain a charging voltage when the battery is charged and a characteristic parameter related to the corresponding discharge capacity; and inputting the obtained characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery. According to the embodiment, the discharge capacity of the battery is predicted according to the extracted characteristic parameters during the charge and discharge of the battery by adopting the pre-trained prediction model, so that the accuracy of prediction of the discharge capacity of the battery is improved.
Drawings
Fig. 1 is a flowchart of a method for predicting battery discharge capacity according to an embodiment of the present disclosure;
FIG. 2 is a flowchart illustrating a method for predicting battery discharge capacity according to an embodiment of the present disclosure;
FIG. 3 is a graph of voltage versus current versus time during a constant current charging phase according to an embodiment of the present application;
FIG. 4 is a graph of discharge capacity with voltage variation after spline interpolation processing in the embodiment of the present application;
FIG. 5 is a graph of a first incremental capacity before smoothing noise reduction according to an embodiment of the present application;
FIG. 6 is a graph comparing a smooth first incremental capacity curve after the smoothing and noise reduction by the Savitzky-Golay filter and a first incremental capacity curve before the smoothing and noise reduction according to the embodiment of the present application;
FIG. 7 is a graph of a smoothed first incremental capacity after smoothing noise reduction according to an embodiment of the present application;
FIG. 8 (a) is a diagram illustrating a feature extracted from a smooth first incremental capacity curve according to an embodiment of the present application;
FIG. 8 (b) is a schematic diagram of another feature extracted from a smooth first incremental capacity curve according to an embodiment of the present application;
FIG. 8 (c) is a schematic diagram of another feature extracted from the smoothed first incremental capacity curve of the application example;
FIG. 8 (d) is a schematic diagram of another feature extracted from the smoothed first incremental capacity curve of the claimed embodiment;
FIG. 8 (e) is a schematic diagram of another feature extracted from a smooth first incremental capacity curve according to an embodiment of the present application;
FIG. 8 (f) is a schematic diagram of another feature extracted from a smoothed first incremental capacity curve according to an embodiment of the present application;
FIG. 8 (g) is a schematic diagram of another feature extracted from a smooth first incremental capacity curve according to an embodiment of the present application;
fig. 9 is a block diagram of a system for predicting battery discharge capacity according to an embodiment of the present disclosure.
Detailed Description
The present invention will be described in further detail with reference to the following detailed description and accompanying drawings. Wherein like elements in different embodiments are numbered with like associated elements. In the following description, numerous details are set forth in order to provide a better understanding of the present application. However, those skilled in the art will readily recognize that some of the features may be omitted or replaced with other elements, materials, methods in different instances. In some instances, certain operations related to the present application have not been shown or described in detail in order to avoid obscuring the core of the present application from excessive description, and it is not necessary for those skilled in the art to describe these operations in detail, so that they may be fully understood from the description in the specification and the general knowledge in the art.
Furthermore, the features, operations, or characteristics described in the specification may be combined in any suitable manner to form various embodiments. Also, the various steps or actions in the method descriptions may be transposed or transposed in order, as will be apparent to one of ordinary skill in the art. Thus, the various sequences in the specification and drawings are for the purpose of clearly describing certain embodiments only and are not intended to imply a required sequence unless otherwise indicated where a certain sequence must be followed.
The ordinal numbers used herein for the components, such as "first," "second," etc., are used merely to distinguish between the objects described, and do not have any sequential or technical meaning. The term "connected" and "coupled" when used in this application, unless otherwise indicated, includes both direct and indirect connections (couplings).
The embodiment provides a method for predicting battery discharge capacity, which includes obtaining a corresponding relation between charging voltage and discharge capacity of a battery, drawing a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharge capacity of the battery according to the corresponding relation, performing smooth noise reduction processing on the first incremental capacity curve by adopting a Savitzky-Golay (least square-based convolution fitting algorithm) filter to obtain a smooth first incremental capacity curve, performing feature extraction on the smooth first incremental capacity curve to obtain feature data of the corresponding relation between the charging voltage and the discharge capacity, inputting the feature data into a pre-trained LSTM (long-short memory) neural network model to obtain residual discharge capacity of the battery, and comparing the collected true capacity with the battery capacity predicted by the method to verify the accuracy of the method for predicting the battery discharge capacity.
Furthermore, after each prediction, the acquired real values of the charging voltage and the discharging capacity of the battery are used as training parameters of the LSTM neural network model, and the LSTM neural network model is further trained and optimized, so that the subsequent prediction result is more accurate.
The first embodiment is as follows:
referring to fig. 1 and fig. 2, the present embodiment provides a method for predicting battery discharge capacity, including:
step 101: acquiring charging voltage and discharging capacity of a plurality of equal interval time points in a battery charging stage;
step 102: processing the charging voltage and the discharging capacity to obtain a plurality of discharging capacities corresponding to the voltage points at equal intervals, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the plurality of voltage points at equal intervals and the discharging capacities corresponding to the voltage points at equal intervals;
step 103: performing characteristic extraction on the first incremental capacity curve to obtain a charging voltage when the battery is charged and a characteristic parameter related to the corresponding discharge capacity;
step 104: and inputting the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery.
The step 101 of obtaining the charging voltage and the discharging capacity at a plurality of time points with equal intervals in the charging stage of the battery includes: and acquiring and recording an equal time interval sequence of the voltage in the constant voltage charging stage from V1 to V2 in the charging period of K previous cycles of the battery and the corresponding charging voltage and discharging capacity, wherein K is more than or equal to 1, and V2 is more than V1. For example, a sequence of equal time intervals when the charging voltage in the constant current charging stage is charged from V1 to V2 in the charging period of k previous cycles of the battery is collected and recorded, wherein the equal time intervals refer to the same time between two collected time points and are marked as { Ti, j | i =1,2, \ 8230;, k, j =1,2, \ 8230;, n }, i represents the number of adopted cyclic charging periods, j represents the serial number of equal time intervals in the sequence of equal time intervals, for example, j =1 represents the first equal interval time interval, and the size of n is determined according to the charging time length of each cycle; simultaneously recording charging voltage, current and discharging capacity data corresponding to each equal time interval point in the equal time interval sequence, for example, the charging is respectively marked as { Vi, j | i =1,2, \ 8230;, k, j =1,2, \ 8230;, n }, { Qi, j | i =1,2, \ 8230;, k, j =1,2, \8230;, n }; collecting battery discharge capacity sequences of k charge and discharge cycles before the lithium battery, and recording the sequences as { Ci | i =1,2,. And k }, wherein the data are pre-used as prediction parameters, namely as input parameters of a neural network model, and in addition, the parameters can also be used as original data of a training data set for training or perfecting the neural network model.
Further, in an embodiment, a sequence of time intervals when the voltage in the constant current charging phase is charged from V1 to V2 in the charging period of N-k cycles after the battery is collected and recorded simultaneously, where N represents the total period of the battery charging and is marked as { Ti, j | i = k +1, k +2, · N, j =1,2, \\ 8230;, N }, and the size of N is determined according to the charging time of each cycle; recording voltage and capacity data corresponding to each equal time interval point in an equal time interval sequence, wherein the data are respectively marked as { Vi, j | i = k +1, k +2,. Sup.,. N, j =1,2, \8230;, N }, { Qi, j | i = k +1, k +2,. Sup.,. N, j =1,2, \8230, N }; the battery discharge capacity sequence of N-k charge and discharge cycles after the lithium battery is collected and is marked as { Ci | i = k +1, k + 2., N }, wherein N is the total number of all charge and discharge cycles, and the data can be used as verification data for verifying whether the prediction method of the application is accurate or not, namely the actually collected discharge capacity can be compared with the discharge capacity predicted by the method of the application to confirm whether the prediction result of the prediction method of the application is accurate or not.
The method includes the steps of performing linear spline interpolation processing on the obtained charging voltage and discharging capacity to obtain discharging capacities corresponding to a plurality of equally spaced voltage points, wherein the equally spaced voltage points refer to the same collected pressure difference between two adjacent voltage points, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the equally spaced voltage points and the corresponding discharging capacities of the equally spaced voltage points specifically includes: the data in the data set is processed, and linear interpolation spline processing is carried out on the voltage sequence { Vi, j | i = m, j =1,2, \ 8230, n } and the capacity sequence { Qi, j | i = m, j =1,2, \ 8230;, n } in the equal time interval data acquired by the mth loop (m ∈ {1,2, \ 8230;, k }). The linear spline interpolation function is composed of n-1 polynomials, the functional expression of the ith polynomial (i =1,2, \8230;, n) is determined by four data points Vm, i, vm, i +1, qm, i +1, and the functional expression is:
wherein any two adjacent polynomials are continuous at the connecting point, and the resulting linear spline interpolation function expression is Q = f (V). The voltage sequence data is divided into equal voltage intervals in the range from V1 to V2, and a first voltage sequence { Vi, j | i = m, j =1,2, \ 8230;, 1000} consisting of 1000 voltages with equal voltage intervals is obtained. Substituting the first voltage sequence { Vi, j | i = m, j =1,2, \ 8230;, 1000} into the linear spline interpolation function expression Q = f (V) yields a first discharging capacity sequence { Qi, j | i = m, j =1,2, \ 8230;, 1000}.
The obtained first voltage series and first discharge capacity series { Vi, j | i = m, j =1,2, \ 8230;, { Qi, j | i = m, j =1,2, \ 8230;, 1000} were subjected to capacity-voltage differential processing to obtain a second voltage series and a second discharge capacity series. Since Vm, j +1-Vm, j is a small value, the difference between two points is used for approximate calculation of the differential, and Δ Q/Δ V is used instead of dQ/dV. Let ICj = Δ Qj/Δ Vj = (Qm, j +1-Qm, j)/(Vm, j +1-Vm, j), the second discharge capacity sequence { ICi, j | i = m, j =1,2, \ 8230;, 999}.
In this embodiment, a Savitzky-Golay filter is used to perform smoothing and denoising processing on the acquired first incremental capacity curve to obtain a smooth curve, which is performed as follows:
in this embodiment, a Savitzky-Golay filter is used, and smooth noise reduction is performed by using the second voltage sequence { Vi, j | i = m, j =1,2, \ 8230;, 999} as an independent variable and the obtained second discharge capacity sequence { ICi, j | i = m, j =1,2, \\ 8230;, 999} as a dependent variable. And selecting five or seven data points as the size of a smoothing window according to the curve smoothing effect and the data loss condition, and selecting a quadratic or cubic polynomial to perform fitting and smoothing. A third discharge capacity sequence { IC x i, j | i = m, j =1,2, \ 8230;, 999} was obtained after smoothing the noise reduction. A smooth first incremental capacity curve after noise reduction is plotted with the second voltage series { Vi, j | i = m, j =1,2, \\ 8230;, 999} as the horizontal axis and the third discharge capacity series { IC _, j | i = m, j =1,2, \ 8230;, 999} as the vertical axis.
It should be noted that, in this embodiment, the differentiation processing and the subsequent smoothing and noise reduction processing are performed on the first incremental capacity curve to further ensure that the extracted feature parameters of the subsequent step are more accurate, so as to ensure that the prediction result is also more accurate.
In step 103, the step of extracting the characteristics of the first incremental capacity curve to obtain the charging voltage when the battery is charged and the characteristic parameters related to the corresponding discharge capacity specifically includes:
in the present embodiment, the obtained smooth first incremental capacity curve is subjected to feature extraction, and the obtained features are determined according to the curve shape, but in the present embodiment, three peaks are obtained according to the curve shape, and the image is divided into three regions by using the local minimum value between the peaks as a boundary point, and the regions are divided as shown in fig. 7. The same features can be extracted from the corresponding positions of each of the 3 blocks. Taking region 2 as an example, 7 features are extracted according to the shape of region 2, and the feature descriptions are shown in fig. 8 (a), 8 (b), 8 (c), 8 (d), 8 (e), 8 (f), and 8 (g). Similarly, corresponding features can be extracted at the same position in the region 1 and the region 3 under the same rule, and 21 features can be extracted in total. Constructing a feature matrix HI using the acquired features m ＝[HI m,1 ,HI m,2 ,…HI m,21 ]。
For example, in this embodiment, a smooth first incremental capacity curve is subjected to feature extraction to obtain a plurality of feature parameters, and the feature parameters form a feature sequence { HIi, j | i = m, j =1,2, \ 8230;, 21}, where j is a feature number, and any one feature parameter in the sequence is a feature matrix.
In this embodiment, the lithium ion battery discharge capacity is predicted based on a long-time and short-time memory LSTM neural network: and training an LSTM model by taking the characteristic parameters extracted from the smooth first incremental capacity curve as input and taking the battery capacity as output, predicting and predicting the future discharge capacity of the lithium ion battery by using the trained LSTM model, realizing online prediction of the discharge capacity of the battery, and outputting the current discharge capacity of the battery after prediction is finished.
In one embodiment, the original data (i.e. the charging voltage and the discharging capacity at the equal interval time point) collected in each cycle of the previous k cycles are processed in the steps 102 to 103, the extracted characteristic parameters are used as training data for training the prediction model, and the training data set added to the prediction model is input into the matrix X train Performing the following steps; processing the original data acquired by each cycle in the last N-k cycles in the steps 102-103, taking the extracted characteristic parameters as verification data, and adding the verification data into an input matrix X of a verification data set of the prediction model test In (1).
In the embodiment, when the prediction model is trained, a feature matrix composed of the charging voltage and the corresponding discharge capacity-related feature parameters is extracted from the first incremental capacity curve, the predicted discharge capacity of the battery is used as output, and the discharge capacity is also expressed in a matrix form. For example in a matrix X train As input, with a matrix Y train As output, an LSTM model training dataset { X is constructed train ，Y train }。X train The matrix is as follows:
Y train the matrix is as follows:
with X test For input, with Y test As output, an LSTM model validation dataset { X is constructed test ，Y test }。X test The matrix is as follows:
Y test the matrix is as follows:
z-score normalization of all the data sets described above, with the matrix X being normalized in units of features train And matrix X test Z-score normalization was performed. I.e. matrix X by column train And matrix X test Z-score normalization was performed. The z-score normalization formula is as follows:
wherein HI is the characteristic value after z-score standardization, HI is the characteristic value before z-score standardization, mu is the column mean value, sigma is the column standard deviation, N is the number of column elements, HI is the number of column elements i Are elements in the column. Respectively obtaining new matrixes X * train And X * test 。X * train The matrix is as follows:
X * test the matrix is as follows:
for matrix Y train And matrix Y test Z-score normalization to obtain a new matrix Y * train And Y * test 。Y * train The matrix is as follows:
Y * test the matrix is as follows:
predicting the discharge capacity of the current cycle, and standardizing the training data set { X ] after z-score treatment * train ，Y * train Inputting the data into an LSTM model for training, and outputting a model training value
Mixing X test Inputting the signal into the trained model to obtain the output->
And the minimum mean square error MSE is used as a model evaluation index for evaluating the model training and predicting effects. With train _ loss = MSE train As an evaluation index of training effect of the training set, the formula is as follows:
with test _ loss = MSE test As an evaluation index of training effect of the training set, the formula is as follows:
and setting different model hyper-parameters Epochs, batch _ size and unit combinations, and training the model. Let in _ loss and test _ loss be observed, and the difference between let in _ loss and test _ loss be observed when both let in _ loss and test _ loss are at lower values. When the difference value between the two is small, the prediction model is not seriously overfitted, and the values of the hyperparameters Epochs, batch _ size and unit can be used in the LSTM model.
Prediction of discharge capacity after i charge-discharge cycle periods, training of data set { X ] after z-score normalization * train ，Y * train In (b), matrix X * train The expression is as follows:
matrix Y * train The expression is as follows:
validation of data set { X) after Z-score normalization * test ，Y * test In the (b) } matrix X * test The expression is as follows:
matrix Y * test The expression is as follows:
and repeating the training steps to train the prediction model and outputting the prediction value of the prediction model.
Further, in another embodiment, the method further comprises the steps of calculating a minimum Mean Square Error (MSE) of a model evaluation index, and adjusting a super-parameter of the model to obtain an optimal model.
Storing the trained prediction model, and performing the operations of step 101 to step 103 on any cycle after N +1 times of charge and discharge cycles of the currently used batteryIn this way, a time span to be predicted is selected, and a feature sequence { HIi, j | i = m, m ] composed of the acquired feature parameters is selected>Inputting N, j =1,2, \8230;, 21} into the trained prediction model, and obtaining the predicted value of the future discharge capacity
Furthermore, when the next online prediction is performed after the N +1 th discharge capacity prediction work is completed, the real battery value of the N +1 th test and the previous N lithium battery state observation data are added into the model data set, and the data set is used for performing a new round of training and prediction on the LSTM model to obtain the predicted value of the N +2 th charge-discharge cycle, so that the online training and prediction are realized.
Compared with the existing prediction method, the prediction method for the battery discharge capacity of the embodiment has the following technical advantages:
1. the embodiment realizes the prediction of the online discharge capacity of the lithium ion battery based on data driving and an LSTM model, and the characteristic parameter identification method is simple, low in complexity and easy to realize in algorithm.
2. The prediction method of the embodiment utilizes the observation data of the battery charging stage to extract the characteristics by a data-driven method, effectively utilizes a large amount of data during the charging and discharging of the battery, and improves the accuracy and robustness of the battery prediction model by continuously training the prediction model.
3. According to the design of the embodiment, the characteristic parameters of the battery during charging and discharging can be extracted on line, the discharging capacity after several future charging and discharging cycles can be predicted every time, the real-time prediction of the future discharging capacity of the battery is realized, and the accuracy of the prediction of the on-line discharging capacity of the battery is improved.
The second embodiment:
in this embodiment, the data source is laboratory accelerated aging experimental data, and the following takes the lithium ion battery operation data shown in table 1 as an example to clearly and completely describe the technical scheme in the embodiment of the present invention.
TABLE 1 accelerated aging test data for lithium ion batteries
In this embodiment, a time interval sequence when the voltage in the constant current charging phase is charged from 2.5V to 3.65V in the charging period of the first 150 cycles of the battery is recorded as { Ti, j | i =1,2, \ 8230;, 150, j =1,2, \ 8230;, n }, where the size of n is determined according to the charging time of each cycle; recording voltages { Vi, j | i =1,2, \ 8230;, 150, j =1,2, \ 8230;, n }, and capacity data discharge capacity sequences { Qi, j | i =1,2, \ 8230;, 150, j =1,2, \ 8230;, n }, corresponding to each equally spaced point in the equally spaced sequence; and collecting battery discharge capacity sequences of the lithium battery for 150 charging and discharging cycles, and recording the sequences as { Ci | i =1, 2., 150}. The time-varying voltage and current of the battery in the constant-current charging stage are shown in fig. 3, and the above data are the original training data for constructing the training data set.
Further, the time interval sequence of charging from 2.5V to 3.65V in the constant current charging stage in the charging period of 50 cycles after the battery is collected and recorded as { T } i，j I =151,152, · 200, j =1,2, \8230;, n }, the size of n depends on the length of each cycle charging time; recording voltage and capacity data corresponding to each equal time interval point in the equal time interval sequence, and respectively recording the voltage and the capacity data as { V i，j |i＝151,152,...,200，j＝1，2，…，n}，{Q i，j I =151,152, ·,200,j =1,2, \8230;, n }; collecting battery discharge capacity sequences of 50 charge-discharge cycles after the lithium battery is charged and marked as { C i I =151,152. The time-varying voltage and current of the battery in the constant-current charging stage are shown in fig. 3, and the data are verification data for verifying the accuracy of the trained prediction model.
The data in the training set is processed by firstly processing the voltage sequence { V in the equal time interval data collected in any cycle m in the first 150 cycles i，j I = m, j =1,2, \ 8230;, n } and capacity sequence { Q | i，j And l i = m, j =1,2, \8230, n } is subjected to linear interpolation spline processing, and the obtained linear spline interpolation function expression is Q = f (V).Dividing the voltage sequence data at equal voltage intervals in the range of 2.5V to 3.65V to obtain 1000 first voltage sequences { V ] with equal voltage intervals i，j I = m, j =1,2, \8230;, 1000}. The first voltage sequence { V } i，j Substituting | i = m, j =1,2, \ 8230;, 1000} into a linear spline interpolation function expression Q = f (V) to obtain a first discharge capacity sequence { Q = f (V) i，j I = m, j =1,2, \ 8230;, 1000}. The first discharge capacity sequence corresponding to the first voltage sequence is shown in fig. 4.
For the first voltage sequence obtained V i，j I = m, j =1,2, \ 8230;, 1000} and a first discharge capacity sequence { Q [ (] i，j The differential processing is carried out to | i = m, j =1,2, \ 8230;, 1000 |, and a second voltage sequence { V | (V) is obtained i，j I = m, j =1,2, \ 8230;, 999 |, and a second discharge capacity sequence { IC } i，j I = m, j =1,2, \8230;, 999}, and the second discharge capacity sequence obtained in this embodiment varies with voltage as shown in fig. 4, which is the first incremental capacity curve before smoothing and noise reduction.
In this embodiment, performing smooth noise reduction processing on the obtained first incremental capacity curve by using a Savitzky-Golay filter, and obtaining a smooth first incremental capacity curve specifically includes: for the second discharge capacity sequence { IC i，j I = m, j =1,2, \ 8230;, 999} smoothing noise reduction using Savitzky-Golay filter with a second voltage sequence { V } i，j I = m, j =1,2, \ 8230;, 999} as an argument to obtain a second discharge capacity sequence { IC i，j I = m, j =1,2, \ 8230;, 999} as a dependent variable, smooth noise reduction is performed. Using seven data points as the smoothing window size, a cubic polynomial is used for fitting and smoothing. Obtaining a third discharge capacity sequence { IC }after smoothing and noise reduction i，j I = m, j =1,2, \ 8230;, 999}. The present embodiment uses the second voltage sequence { V } i，j I = m, j =1,2, \ 8230;, 999} is plotted as the horizontal axis with the third discharge capacity sequence { IC after noise reduction * i，j I = m, j =1,2, \ 8230, 999} as the vertical axis, a smooth first incremental capacity curve is plotted. The smooth first incremental capacity curve after the noise reduction processing in this embodiment is shown in fig. 6.
In this embodiment, the performing feature extraction on the obtained smooth first incremental capacity curve includes: three peaks are obtained from the smooth first incremental capacity curve shape, and the image is divided into three regions with the local minimum between the peaks as a boundary point, and the region division is shown in fig. 7. The same features can be extracted from the corresponding positions of each of the 3 blocks. Taking region 2 as an example, 7 features are extracted according to the shape of region 2, and the description of the features is shown in fig. 8. Similarly, corresponding features can be extracted at the same position in the region 1 and the region 3 under the same rule, and 21 features can be extracted in total. Constructing a feature matrix HI using the acquired features m ＝[HI m,1 ,HI m,2 ,…HI m,21 ]。
In this embodiment, the discharge capacity of the lithium ion battery is predicted based on a long-time and short-time memory LSTM neural network. The method specifically comprises the following steps: the characteristic parameters obtained after the above processing is carried out on the original data collected in each cycle of the previous 150 cycles are added into a training data set input matrix X of the prediction model train The discharge capacity of the previous 150 cycles of charge and discharge cycles was added to the matrix Y train (ii) a The original data collected in each cycle of the last 50 cycles are processed by the method so as to obtain characteristic parameters for verifying the prediction model, and the characteristic parameters for verifying the prediction model are added into a verification data set input matrix X of the prediction model test The discharge capacity of the last 50 cycles of charge and discharge cycles was added to the matrix Y test 。
In matrix X train As input, with a matrix Y train As output, training the LSTM model training dataset { X train ，Y train }. With X test For input, with Y test As output, an LSTM model validation dataset { X is constructed test ，Y test }. All data sets were subjected to z-score normalization.
The training data set after z-score standardization is processed by X * train ，Y * train Inputting the data into LSTM model for training, and outputting model training value
Mixing X test Inputting the signal into a trained prediction model to obtain a model output->
And the minimum mean square error MSE is used as a model evaluation index for evaluating the model training and predicting effects.
And setting different model hyper-parameters Epochs, batch _ size and unit combination to train the model. When train _ loss = MSE train And test _ loss = MSE test Both at lower values and when the difference between the two is small (typically less than 0.05), the values of the hyperparameters Epochs, batch _ size, unit are used in the LSTM model.
In this example, discharge capacities after 1, 3, 5, 8, and 10 charge-discharge cycle periods were predicted. In the embodiment, after prediction is finished, the acquired real data is used for further training a prediction model, different model super-parameter combinations are selected and input into the model in the model training process, the model is trained to obtain a training result, and the minimum Mean Square Error (MSE) of a model evaluation index is calculated; and selecting the group of model hyper-parameters corresponding to the MSE minimum value in each result, and selecting the group of hyper-parameter input models to obtain the optimal model.
The model hyper-parameter settings of the prediction model of the present embodiment are as follows: number of neuron layers =2; epochs =32; batch _ size =1; first tier unit =1; second tier unit =4; optimizer = adam.
Processing any cycle after 250 times of charge and discharge cycles of the currently used battery to obtain a characteristic sequence, and acquiring the characteristic sequence { HI } i，j |i＝m，m>N, j =1,2, \ 8230;, 21} is input into a trained prediction model to obtain a predicted value of the future discharge capacity. When the prediction work of the next cycle is needed after the discharge capacity prediction work after the 251 th cycle is finished, the true battery value of the 251 th cycle and the previous 250 lithium battery state observation data are added into a model data set, and the data set is used for carrying out a new round of training and prediction on an LSTM model to obtain the 252 th chargingAnd the predicted value of the discharge cycle period realizes on-line training and prediction.
The prediction model training and prediction results of this example are shown in table 2. As can be seen from table 2, in the present embodiment, the battery charging data and the historical data are fully utilized, the discharging capacity after n times of charging and discharging cycles can be predicted each time, and the model can be trained and predicted on line to determine the future state of health change of the battery. As shown in table 2, when the discharge capacity within 5 cycles is predicted, the error of the model training set can be controlled within 0.02, and the error of the verification set can be controlled within 0.05; when the discharge capacity within 10 cycles is predicted, the error of a model training set can be controlled within 0.04, and the error of a verification set can be controlled within 0.09; the model of the invention has higher prediction precision, can predict the discharge capacity within ten charge-discharge cycle periods, and provides a basis for judging the recent change condition of the health state of the lithium battery.
TABLE 2 model training and prediction results
EXAMPLE III
The embodiment provides a system for predicting battery discharge capacity, which comprises an obtaining module 201, a first processing module 202, an extracting module 203 and a predicting module 204, wherein the obtaining module 201 is used for obtaining the charge voltage and the discharge capacity of a plurality of time points with equal intervals in a battery charging stage; the first processing module 202 is configured to perform linear spline interpolation processing on the obtained charging voltage and discharging capacity to obtain discharging capacities corresponding to a plurality of equally spaced voltage points, and obtain a first incremental capacity curve used for representing a corresponding relationship between the charging voltage and the discharging capacity of the battery according to the plurality of equally spaced voltage points and the discharging capacities corresponding to the equally spaced voltage points; the extraction module 203 is configured to perform feature extraction on the first incremental capacity curve to obtain a charging voltage when the battery is charged and a feature parameter related to a corresponding discharge capacity; the prediction module 204 is configured to input the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery.
In another embodiment, the system for predicting battery discharge capacity further includes an output module 205, for example, the output module is a display screen, and the output module is configured to output and display the predicted current discharge capacity of the battery, so that a worker can visually check the predicted current discharge capacity of the battery conveniently.
The specific predicted operating strategy in each module is the same as that in the first embodiment, and is not described herein again.
Example four
The present embodiment provides a computer-readable storage medium including a program, the program being executable by a processor to implement the battery discharge capacity prediction method as provided in the first embodiment.
The present invention has been described in terms of specific examples, which are provided to aid in understanding the invention and are not intended to be limiting. For a person skilled in the art to which the invention pertains, several simple deductions, modifications or substitutions may be made according to the idea of the invention.
Claims (5)
1. A method for predicting battery discharge capacity, comprising:
acquiring charging voltage and discharging capacity of a plurality of equal interval time points in a battery charging stage;
processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of equally spaced voltage points, and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacities of the battery according to the equally spaced voltage points and the corresponding discharging capacities thereof;
extracting the characteristics of the first incremental capacity curve to obtain the charging voltage when the battery is charged and the corresponding characteristic parameters related to the discharge capacity;
inputting the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery;
after obtaining the discharge capacity corresponding to a plurality of equally spaced voltage points, the method also comprises the following steps: carrying out differential processing of discharge capacity relative to charge voltage on the plurality of voltage points and the corresponding discharge capacities thereof;
after obtaining the first incremental capacity curve, the method further comprises the following steps: performing smooth noise reduction processing on the first incremental capacity curve by adopting a Savitzky-Golay filter;
the acquiring of the charging voltage and the discharging capacity of a plurality of equal interval time points in the charging stage of the battery comprises the following steps:
acquiring and recording an equal time interval sequence of a constant voltage charging stage in a charging period of K previous cycles of the battery from V1 to V2 and corresponding charging voltage and discharging capacity, wherein the equal time interval sequence comprises a plurality of equal interval time points, K is more than or equal to 1, and V2 is more than V1;
the step of processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of voltage points with equal intervals comprises the following steps:
dividing a voltage interval consisting of charging voltages V1-V2 into a first voltage sequence comprising a plurality of voltage points at equal intervals, and collecting the discharge capacity corresponding to each voltage point at equal intervals;
the differential processing of the discharge capacity with respect to the charge voltage on the plurality of voltage points and the corresponding discharge capacities thereof comprises:
substituting the first voltage sequence into a preset linear spline interpolation function expression to obtain a first discharge capacity sequence;
carrying out differential processing of the discharge capacity relative to the charging voltage on the first discharge capacity sequence and the first voltage sequence to obtain a second voltage sequence and a second discharge capacity sequence corresponding to the second voltage sequence;
the obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the plurality of equally spaced voltage points and the corresponding discharging capacities thereof comprises: drawing a first incremental capacity curve by taking the second voltage sequence as a horizontal axis and the second discharge capacity sequence as a vertical axis;
the smoothing and denoising the first incremental capacity curve by using the Savitzky-Golay filter comprises:
and adopting a Savitzky-Golay filter to perform smooth noise reduction treatment by taking the second voltage sequence as an independent variable and the second discharge capacity sequence as a dependent variable to obtain a third discharge capacity sequence, and drawing a smooth first incremental capacity curve by taking the second voltage sequence as a horizontal axis and the third discharge capacity sequence as a vertical axis.
2. The method for predicting the discharge capacity of the battery according to claim 1, wherein after the characteristic parameter is input into a pre-trained prediction model to obtain the current discharge capacity of the battery, the method further comprises the step of taking the characteristic parameter as training data of the prediction model to further train the prediction model.
3. The method for predicting battery discharge capacity according to claim 1, wherein a minimum mean square error of the prediction model is calculated during training of the prediction model, and the hyper-parameters of the prediction model are adjusted according to the minimum mean square error to optimize the prediction model.
4. A battery discharge capacity prediction system for applying the battery discharge capacity prediction method according to any one of claims 1 to 3, the battery discharge capacity prediction system comprising:
the acquisition module is used for acquiring the charging voltage and the discharging capacity of a plurality of equal interval time points in the charging stage of the battery;
the first processing module is used for processing the charging voltage and the discharging capacitor to obtain discharging capacities corresponding to a plurality of equally spaced voltage points and obtaining a first incremental capacity curve for representing the corresponding relation between the charging voltage and the discharging capacity of the battery according to the equally spaced voltage points and the corresponding discharging capacities thereof;
the extraction module is used for extracting the characteristics of the first incremental capacity curve to obtain the charging voltage when the battery is charged and the corresponding characteristic parameters related to the discharge capacity;
and the prediction module is used for inputting the characteristic parameters into a pre-trained prediction model to obtain the current discharge capacity of the battery.
5. A computer-readable storage medium, characterized by comprising a program executable by a processor to implement the method of any one of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010960776.3A CN112098848B (en) | 2020-09-14 | 2020-09-14 | Battery discharge capacity prediction method and system and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010960776.3A CN112098848B (en) | 2020-09-14 | 2020-09-14 | Battery discharge capacity prediction method and system and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112098848A CN112098848A (en) | 2020-12-18 |
| CN112098848B true CN112098848B (en) | 2023-04-11 |
Family
ID=73751512
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010960776.3A Active CN112098848B (en) | 2020-09-14 | 2020-09-14 | Battery discharge capacity prediction method and system and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112098848B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113049974B (en) * | 2021-02-02 | 2023-01-13 | 欣旺达电动汽车电池有限公司 | Battery capacity prediction method, device, equipment and storage medium |
| CN113447817B (en) * | 2021-05-13 | 2022-04-08 | 浙江大学 | Lithium battery online life prediction method based on two-point life characteristics |
| CN113687250B (en) * | 2021-08-18 | 2023-10-24 | 蜂巢能源科技有限公司 | Battery cell capacity prediction method and device, electronic equipment and medium |
| CN113820608B (en) * | 2021-08-20 | 2022-08-05 | 北京邮电大学 | Method for predicting remaining capacity of battery in echelon and electronic equipment |
| CN113671383B (en) * | 2021-09-15 | 2023-05-26 | 中国计量大学 | Lithium ion battery residual life prediction system and control method thereof |
| CN114859249B (en) * | 2022-07-06 | 2022-10-18 | 苏州清研精准汽车科技有限公司 | Method and device for detecting battery pack capacity |
| CN115656834A (en) * | 2022-11-02 | 2023-01-31 | 武汉动力电池再生技术有限公司 | Battery capacity prediction method and device and electronic equipment |
| CN116449209A (en) * | 2023-01-12 | 2023-07-18 | 帕诺(常熟)新能源科技有限公司 | Actual operation energy storage lithium capacitance prediction method based on LSTM |
| CN115993538B (en) * | 2023-02-01 | 2023-10-20 | 上海玫克生储能科技有限公司 | Fitting method and device of battery capacity increment comprehensive curve and electronic equipment |
| CN116027204B (en) * | 2023-02-20 | 2023-06-20 | 山东大学 | Lithium battery residual service life prediction method and device based on data fusion |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103389471A (en) * | 2013-07-25 | 2013-11-13 | 哈尔滨工业大学 | Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register) |
| CN109031153A (en) * | 2018-10-16 | 2018-12-18 | 北京交通大学 | A kind of health status On-line Estimation method of lithium ion battery |
| CN110568361A (en) * | 2019-09-12 | 2019-12-13 | 华中科技大学 | Method for predicting health state of power battery |
| CN110703101A (en) * | 2019-09-12 | 2020-01-17 | 北京交通大学 | Lithium ion battery inter-partition cycle capacity decline prediction method |
| CN111090047A (en) * | 2019-12-09 | 2020-05-01 | 泉州装备制造研究所 | Lithium battery health state estimation method based on multi-model fusion |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107656210B (en) * | 2017-09-14 | 2020-01-21 | 广州市香港科大霍英东研究院 | Method for estimating battery electric quantity state |
-
2020
- 2020-09-14 CN CN202010960776.3A patent/CN112098848B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103389471A (en) * | 2013-07-25 | 2013-11-13 | 哈尔滨工业大学 | Cycle life indirect prediction method for lithium ion battery provided with uncertain intervals on basis of GPR (general purpose register) |
| CN109031153A (en) * | 2018-10-16 | 2018-12-18 | 北京交通大学 | A kind of health status On-line Estimation method of lithium ion battery |
| CN110568361A (en) * | 2019-09-12 | 2019-12-13 | 华中科技大学 | Method for predicting health state of power battery |
| CN110703101A (en) * | 2019-09-12 | 2020-01-17 | 北京交通大学 | Lithium ion battery inter-partition cycle capacity decline prediction method |
| CN111090047A (en) * | 2019-12-09 | 2020-05-01 | 泉州装备制造研究所 | Lithium battery health state estimation method based on multi-model fusion |
Non-Patent Citations (1)
| Title |
|---|
| 任鹏博.基于区间应力的锂离子电池容量估计及寿命预测研究.《中国优秀硕士学位论文全文数据库》.2020,(第1期), * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112098848A (en) | 2020-12-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112098848B (en) | Battery discharge capacity prediction method and system and readable storage medium | |
| CN110068774B (en) | Lithium battery health state estimation method and device and storage medium | |
| CN111443294B (en) | Method and device for indirectly predicting remaining life of lithium ion battery | |
| CN112540317B (en) | Battery health state estimation and residual life prediction method based on real vehicle data | |
| CN109978229A (en) | The method that the full battery core multi-point temperature of a kind of pair of power battery pack and tie point temperature carry out thermal runaway prediction | |
| WO2014054259A1 (en) | Device for estimating parameters of battery, and estimation method | |
| CN108287317B (en) | Battery power prediction model generation method and system and power prediction method and system | |
| CN103399280B (en) | Based on the cycle life of lithium ion battery Forecasting Methodology of NSDP-AR model | |
| CN112904219B (en) | Big data-based power battery health state prediction method | |
| CN110568360B (en) | Lithium battery aging diagnosis method based on fuzzy logic algorithm | |
| Huang et al. | A comprehensively optimized lithium-ion battery state-of-health estimator based on Local Coulomb Counting Curve | |
| CN111157908A (en) | Battery health degree prediction method based on big data machine learning | |
| CN113406500B (en) | Method for estimating residual electric quantity of power lithium battery | |
| CN115201686B (en) | Lithium ion battery health state assessment method under incomplete charge and discharge data | |
| CN114290960A (en) | Method and device for acquiring battery health degree of power battery and vehicle | |
| CN113406503A (en) | Lithium battery SOH online estimation method based on deep neural network | |
| CN115586444A (en) | Lithium battery residual life prediction method based on VMD and BP neural network | |
| CN115267575A (en) | Energy storage battery life prediction algorithm based on Transformer | |
| DE102018108738A1 (en) | Method for determining a state of aging of a battery, computer program, storage means, control unit and vehicle | |
| CN116736130A (en) | Lithium battery residual service life prediction method and system | |
| CN115219907A (en) | Lithium battery SOC estimation method, system, medium, equipment and terminal | |
| CN113255215B (en) | Lithium battery health state estimation method based on voltage segments | |
| CN114545276A (en) | Power battery service life prediction method based on capacity test and Internet of vehicles big data | |
| CN115980607A (en) | Battery state of health (SOH) estimation method based on incremental capacity curve characteristic points | |
| CN117330987B (en) | Method, system, medium and apparatus for time-based battery state of health assessment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |